BZPEER
More

Discover

About BZPEERSearchGraph

Research tools

HomeCollectionsResearchCollaborationPublish

Network

ExpertsCommunities

Opportunities

Listings

Account

Log inCreate account
Log inCreate account

Source author record

Xing Fu

Explore this author’s publications, collaborators and research topics. This catalog record has not been claimed by its author.

ResearcherUnclaimed source record
physics.opticsmath.CAmath.FAArtificial IntelligenceComputation and LanguageComputer Visioneess.IVInformation RetrievalMachine Learningmath.DSphysics.med-phquant-ph

Catalog footprint

Research overview

16works
12topics
4close collaborators

Actions

Is this your work?

Log in to claim
Open graphBrowse papers

Published work

16 publications

preprint2026arXiv

1.1 kW, 100 Hz room-temperature diode-pumped nanosecond laser by water immersion cooling

We report a room-temperature diode-pumped solid-state laser by water immersion cooling, which delivers a pulse energy of 11 J at the repetition rate of 100 Hz and the pulse duration of 7 ns, while the beam quality factor is 2.6 times the diffraction limit. To the best of our knowledge, this represents the highest performance achieved for room-temperature nanosecond lasers operating above 100 Hz, which demonstrates the great potentials of room-temperature immersion-cooled nanosecond active mirror lasers.

Open access
preprint2026arXiv

Multivariate Diffusion Transformer with Decoupled Attention for High-Fidelity Mask-Text Collaborative Facial Generation

While significant progress has been achieved in multimodal facial generation using semantic masks and textual descriptions, conventional feature fusion approaches often fail to enable effective cross-modal interactions, thereby leading to suboptimal generation outcomes. To address this challenge, we introduce MDiTFace--a customized diffusion transformer framework that employs a unified tokenization strategy to process semantic mask and text inputs, eliminating discrepancies between heterogeneous modality representations. The framework facilitates comprehensive multimodal feature interaction through stacked, newly designed multivariate transformer blocks that process all conditions synchronously. Additionally, we design a novel decoupled attention mechanism by dissociating implicit dependencies between mask tokens and temporal embeddings. This mechanism segregates internal computations into dynamic and static pathways, enabling caching and reuse of features computed in static pathways after initial calculation, thereby reducing additional computational overhead introduced by mask condition by over 94% while maintaining performance. Extensive experiments demonstrate that MDiTFace significantly outperforms other competing methods in terms of both facial fidelity and conditional consistency.

Open access
preprint2026arXiv

TabEmbed: Benchmarking and Learning Generalist Embeddings for Tabular Understanding

Foundation models have established unified representations for natural language processing, yet this paradigm remains largely unexplored for tabular data. Existing methods face fundamental limitations: LLM-based approaches lack retrieval-compatible vector outputs, whereas text embedding models often fail to capture tabular structure and numerical semantics. To bridge this gap, we first introduce the Tabular Embedding Benchmark (TabBench), a comprehensive suite designed to evaluate the tabular understanding capability of embedding models. We then propose TabEmbed, the first generalist embedding model that unifies tabular classification and retrieval within a shared embedding space. By reformulating diverse tabular tasks as semantic matching problems, TabEmbed leverages large-scale contrastive learning with positive-aware hard negative mining to discern fine-grained structural and numerical nuances. Experimental results on TabBench demonstrate that TabEmbed significantly outperforms state-of-the-art text embedding models, establishing a new baseline for universal tabular representation learning. Code and datasets are publicly available at https://github.com/qiangminjie27/TabEmbed and https://huggingface.co/datasets/qiangminjie27/TabBench.

Open access
preprint2026arXiv

Table as a Modality for Large Language Models

To migrate the remarkable successes of Large Language Models (LLMs), the community has made numerous efforts to generalize them to the table reasoning tasks for the widely deployed tabular data. Despite that, in this work, by showing a probing experiment on our proposed StructQA benchmark, we postulate that even the most advanced LLMs (such as GPTs) may still fall short of coping with tabular data. More specifically, the current scheme often simply relies on serializing the tabular data, together with the meta information, then inputting them through the LLMs. We argue that the loss of structural information is the root of this shortcoming. In this work, we further propose TAMO, which bears an ideology to treat the tables as an independent modality integrated with the text tokens. The resulting model in TAMO is a multimodal framework consisting of a hypergraph neural network as the global table encoder seamlessly integrated with the mainstream LLM. Empirical results on various benchmarking datasets, including HiTab, WikiTQ, WikiSQL, FeTaQA, and StructQA, have demonstrated significant improvements on generalization with an average relative gain of 42.65%.

Open access
preprint2022arXiv

3D inhomogeneous self-accelerating beams

We propose and generate a new class of structured light fulfilling quantum-like coherent states based on a set of circular Airy vortex modes. Such coherent-state wave packets possess strong focus with both radial and angular self-accelerations, which exploit more general 3D inhomogeneous velocity control with global spatial symmetry of multilayer rotation akin to galactic kinematics, as termed galaxy waves. Galaxy waves are endowed with new degrees of freedom to control strong focusing and acceleration of 3D structured light, promising numerous applications in optical trapping, manufacturing, and nonlinear optics.

Open access
preprint2022arXiv

Particle manipulation behind turbid medium based on intensity transmission matrix

Optical tweezers can manipulate tiny particles. However, the distortion caused by the scattering medium restricts the applications of optical tweezers. Wavefront shaping techniques including the transmission matrix (TM) method are powerful tools to achieve light focusing behind the scattering medium. In this paper, we propose a new kind of TM, named intensity transmission matrix (ITM). Only relying on the intensity distribution, we can calculate the ITM with only about 1/4 measurement time of the widely used four-phase method. Meanwhile, ITM method can avoid the energy loss in diffraction introduced by holographic modulation. Based on the ITM, we have implemented particle manipulation with a high degree of freedom on single and multiple particles. In addition, the manipulation range is enlarged over twenty times (compared with the memory effect) to 200 μm.

Open access
preprint2021arXiv

Coherent ray-wave structured light based on (helical) Ince-Gaussian modes

The topological evolution of classic eigenmodes including Hermite-Laguerre-Gaussian and (helical) InceGaussian modes is exploited to construct coherent state modes, which unifies the representations of travelingwave (TW) and standing-wave (SW) ray-wave structured light for the first time and realizes the TW-SW unified ray-wave geometric beam with topology of raytrajectories splitting effect, breaking the boundary of TW and SW structured light. We experimentally generate these new modes with high purity and dynamic control by digital holography method, revealing potential applications in optical manipulation and communication.

Open access
preprint2020arXiv

A One-Shot Learning Framework for Assessment of Fibrillar Collagen from Second Harmonic Generation Images of an Infarcted Myocardium

Myocardial infarction (MI) is a scientific term that refers to heart attack. In this study, we infer highly relevant second harmonic generation (SHG) cues from collagen fibers exhibiting highly non-centrosymmetric assembly together with two-photon excited cellular autofluorescence in infarcted mouse heart to quantitatively probe fibrosis, especially targeted at an early stage after MI. We present a robust one-shot machine learning algorithm that enables determination of 2D assembly of collagen with high spatial resolution along with its structural arrangement in heart tissues post-MI with spectral specificity and sensitivity. Detection, evaluation, and precise quantification of fibrosis extent at early stage would guide one to develop treatment therapies that may prevent further progression and determine heart transplant needs for patient survival.

Open access
preprint2020arXiv

High-dimensional classically entangled light from a laser

Vectorially structured light has emerged as an enabling tool in many diverse applications, from communication to imaging, exploiting quantum-like correlations courtesy of a non-separable spatially varying polarization structure. Creating these states at the source remains challenging and is presently limited to two-dimensional vectorial states by customized lasers. Here we invoke ray-wave duality in a simple laser cavity to produce polarization marked multi-path modes that are non-separable in three degrees of freedom and in eight dimensions. As a topical example, we use our laser to produce the complete set of Greenberger-Horne-Zeilinger (GHZ) basis states, mimicking high-dimensional multi-partite entanglement with classical light, which we confirm by a new projection approach. We offer a complete theoretical framework for our laser based on SU(2) symmetry groups, revealing a rich parameter space for further exploitation. Our approach requires only a conventional laser with no special optical elements, is easily scaleable to higher dimensions, and offers a simple but elegant solution for at-the-source creation of classically entangled states of structured light, opening new applications in simulating and enhancing high-dimensional quantum systems.

Open access
preprint2020arXiv

Two-dimensional-controlled high-order modes and vortex beams from an intracavity mode converter laser

We present a novel scheme of structured light laser with an astigmatic mode converter (AMC) as intracavity element, first enabling the generation of Hermite-Gaussian (HG) modes with fully controlled two-dimensional (2D) indices (m,n) and vortex beams carrying orbital angular momentum (OAM) directly from cavity. The 2D tunability was realized by controlling the off-axis displacements of both pump and intracavity AMC. The output HGm,n beam could be externally converted into OAM beam with 2D tunable radial and azimuthal indices (p,l). With the certain parameter control, vortex beam carrying OAM also could be directly generated from the cavity. Our setup provides a compact and concise structured light source. It has great potential in extending various applications of optical tweezers, communications, and nonlinearity.

Open access
preprint2016arXiv

Riesz Transform Characterization and Fefferman-Stein Decomposition of Triebel-Lizorkin Spaces

Let D∈ND\in\mathbb{N}, q∈[2,∞)q\in[2,\infty) and (RD,∣⋅∣,dx)(\mathbb{R}^D,|\cdot|,dx) be the Euclidean space equipped with the DD-dimensional Lebesgue measure. In this article, via an auxiliary function space WE1, q(RD)\mathrm{WE}^{1,\,q}(\mathbb R^D) defined via wavelet expansions, the authors establish the Riesz transform characterization of Triebel-Lizorkin spaces F˙1, q0(RD)\dot{F}^0_{1,\,q}(\mathbb{R}^D). As a consequence, the authors obtain the Fefferman-Stein decomposition of Triebel-Lizorkin spaces F˙∞, q′0(RD)\dot{F}^0_{\infty,\,q'}(\mathbb{R}^D). Finally, the authors give an explicit example to show that F˙1, q0(RD)\dot{F}^0_{1,\,q}(\mathbb{R}^D) is strictly contained in WE1, q(RD)\mathrm{WE}^{1,\,q}(\mathbb{R}^D) and, by duality, WE∞, q′(RD)\mathrm{WE}^{\infty,\,q'}(\mathbb{R}^D) is strictly contained in F˙∞, q′0(RD)\dot{F}^0_{\infty,\,q'}(\mathbb{R}^D). Although all results when D=1D=1 were obtained by C.-C. Lin et al. [Michigan Math. J. 62 (2013), 691-703], as was pointed out by C.-C. Lin et al., the approach used in the case D=1D=1 can not be applied to the case D≥2D\ge2, which needs some new skills.

Open access
preprint2015arXiv

Multi-Resolution Dynamic Mode Decomposition

We demonstrate that the integration of the recently developed dynamic mode decomposition (DMD) with a multi-resolution analysis allows for a decomposition method capable of robustly separating complex systems into a hierarchy of multi-resolution time-scale components. A one-level separation allows for background (low-rank) and foreground (sparse) separation of dynamical data, or robust principal component analysis. The multi-resolution dynamic mode decomposition is capable of characterizing nonlinear dynamical systems in an equation-free manner by recursively decomposing the state of the system into low-rank terms whose temporal coefficients in time are known. DMD modes with temporal frequencies near the origin (zero-modes) are interpreted as background (low-rank) portions of the given dynamics, and the terms with temporal frequencies bounded away from the origin are their sparse counterparts. The multi-resolution dynamic mode decomposition (mrDMD) method is demonstrated on several examples involving multi-scale dynamical data, showing excellent decomposition results, including sifting the El Niño mode from ocean temperature data. It is further applied to decompose a video data set into separate objects moving at different rates against a slowly varying background. These examples show that the decomposition is an effective dynamical systems tool for data-driven discovery.

Open access
preprint2015arXiv

Products of Functions in BMO(X){\mathop\mathrm{BMO}}({\mathcal X}) and Hat1(X)H^1_{\rm at}({\mathcal X}) via Wavelets over Spaces of Homogeneous Type

Let (X,d,μ)({\mathcal X},d,μ) be a metric measure space of homogeneous type in the sense of R. R. Coifman and G. Weiss and Hat1(X)H^1_{\rm at}({\mathcal X}) be the atomic Hardy space. Via orthonormal bases of regular wavelets and spline functions recently constructed by P. Auscher and T. Hytönen, the authors prove that the product f×gf\times g of f∈Hat1(X)f\in H^1_{\rm at}({\mathcal X}) and g∈BMO(X)g\in\mathop\mathrm{BMO}({\mathcal X}), viewed as a distribution, can be written into a sum of two bounded bilinear operators, respectively, from Hat1(X)×BMO(X)H^1_{\rm at}({\mathcal X})\times\mathop\mathrm{BMO}({\mathcal X}) into L1(X)L^1({\mathcal X}) and from Hat1(X)×BMO(X)H^1_{\rm at}({\mathcal X})\times\mathop\mathrm{BMO}({\mathcal X}) into Hlog⁡(X)H^{\log}({\mathcal X}), which affirmatively confirms the conjecture suggested by A. Bonami and F. Bernicot (This conjecture was presented by L. D. Ky in [J. Math. Anal. Appl. 425 (2015), 807-817]).

Open access
preprint2015arXiv

Wavelet Characterizations of the Atomic Hardy Space H1H^1 on Spaces of Homogeneous Type

Let (X,d,μ)({\mathcal X},d,μ) be a metric measure space of homogeneous type in the sense of R. R. Coifman and G. Weiss and Hat1(X)H^1_{\rm at}({\mathcal X}) be the atomic Hardy space. Via orthonormal bases of regular wavelets and spline functions recently constructed by P. Auscher and T. Hytönen, together with obtaining some crucial lower bounds for regular wavelets, the authors give an unconditional basis of Hat1(X)H^1_{\rm at}({\mathcal X}) and several equivalent characterizations of Hat1(X)H^1_{\rm at}({\mathcal X}) in terms of wavelets, which are proved useful.

Open access
preprint2014arXiv

Generalized Fractional Integrals and Their Commutators over Non-homogeneous Metric Measure Spaces

Let (X,d,μ)({\mathcal X},d,μ) be a metric measure space satisfying both the upper doubling and the geometrically doubling conditions. In this paper, the authors establish some equivalent characterizations for the boundedness of fractional integrals over (X,d,μ)({\mathcal X},d,μ). The authors also prove that multilinear commutators of fractional integrals with  RBMO(μ){\mathop\mathrm{\,RBMO(μ)}} functions are bounded on Orlicz spaces over (X,d,μ)({\mathcal X},d,μ), which include Lebesgue spaces as special cases. The weak type endpoint estimates for multilinear commutators of fractional integrals with functions in the Orlicz-type space Oscexp⁡Lr(μ){\mathrm{Osc}_{\exp L^r}(μ)}, where r∈[1,∞)r\in [1,\infty), are also presented. Finally, all these results are applied to a specific example of fractional integrals over non-homogeneous metric measure spaces.

Open access
preprint2014arXiv

Hardy spaces HpH^p over non-homogeneous metric measure spaces and their applications

Let (X,d,μ)({\mathcal X},d,μ) be a metric measure space satisfying both the geometrically doubling and the upper doubling conditions. Let ρ∈(1,∞)ρ\in (1,\infty), 0<p≤1≤q≤∞0<p\le1\le q\le\infty, p≠qp\neq q, γ∈[1,∞)γ\in[1,\infty) and ε∈(0,∞)ε\in(0,\infty). In this article, the authors introduce the atomic Hardy space H~atb, ρp, q, γ(μ){\widetilde H_{\mathrm{atb},\,ρ}^{p,\,q,\,γ}(μ)} and the molecular Hardy space H~mb, ρp, q, γ, ε(μ){\widetilde H_{\rm{mb},\,ρ}^{p,\,q,\,γ,\,ε}(μ)} via the discrete coefficient K~B, S(ρ), p\widetilde{K}^{(ρ),\,p}_{B,\,S}, and prove that the Calderón-Zygmund operator is bounded from H~mb, ρp, q, γ, δ(μ){\widetilde H_{\rm{mb},\,ρ}^{p,\,q,\,γ,\,δ}(μ)} (or H~atb, ρp, q, γ(μ){\widetilde H_{\rm{atb},\,ρ}^{p,\,q,\,γ}(μ)}) into Lp(μ)L^p(μ), and from H~atb, ρ(ρ+1)p, q, γ+1(μ){\widetilde H_{\rm{atb},\,ρ(ρ+1)}^{p,\,q,\,γ+1}(μ)} into H~mb, ρp, q, γ, 12(δ−νp+ν)(μ){\widetilde H_{\rm{mb},\,ρ}^{p,\,q,\,γ,\,\frac12(δ-\fracν{p}+ν)}(μ)} whose genealized fractional versions are also obtained. The authors also introduce the ρρ-weakly doubling condition, with ρ∈(1,∞)ρ\in (1,\infty), of the measure μμ and construct a non-doubling measure μμ satisfying this condition. If μμ is ρρ-weakly doubling, the authors further introduce the Campanato space Eρ, η, γα, q(μ){\mathcal E}^{α,\,q}_{ρ,\,η,\,γ}(μ) and show that Eρ, η, γα, q(μ){\mathcal E}^{α,\,q}_{ρ,\,η,\,γ}(μ) is independent of the choices of ρρ, ηη, γγ and qq; the authors then introduce the atomic Hardy space H^atb, ρp, q, γ(μ)\widehat H_{\rm{atb},\,ρ}^{p,\,q,\,γ}(μ) and the molecular Hardy space H^mb, ρp, q, γ, ε(μ)\widehat H_{\rm{mb},\,ρ}^{p,\,q,\,γ,\,ε}(μ), which coincide with each other; the authors finally prove that H^atb, ρp, q, γ(μ)\widehat{H}_{\rm{atb},\,ρ}^{p,\,q,\,γ}(μ) is the predual of Eρ, ρ, 11/p−1, 1(μ){\mathcal E}^{1/p-1,\,1}_{ρ,\,ρ,\,1}(μ). Moreover, relations of these Hardy spaces are also discussed.

Open access

Research graph

See the researcher in context

Open full explorer

Inspect adjacent papers, topics, institutions and collaborators without losing the researcher page.

Building this map preview

BZPEER is loading the nearby papers, people, topics and institutions for this page.